Which of the following models is the backward push method?

The following is the back propagation of the neural network model.

The backstepping method of neural network model is a method to update the weights of neural network through back propagation error. It is a supervised learning method, and its basic idea is to transfer an input sample from the input layer to the output layer, then calculate the error of the output layer, and propagate the error back to each neuron, thus updating the weight.

Specifically, the process of the backward extrapolation method includes the following steps:

1. Forward propagation: the input value is sent to the input layer, and the output value is obtained through multi-layer calculation.

2. Calculation error: compare the output value with the actual value and calculate the error.

3. Back propagation: the error propagates back from the output layer to the input layer, and the error contribution of each layer is calculated by the chain rule.

4. Update weights: update the weights of each neuron according to the contribution of errors and the learning rate.

5. Repeat step 1-4 until the predefined training times or error range is reached.

Backstepping can update the weights of neural network through back propagation error, which can effectively improve the accuracy and generalization ability of neural network model, so it is widely used in image recognition, natural language processing, speech recognition and other fields.